Power, Politics, and the Planetary Costs of Artificial Intelligence
Total Page:16
File Type:pdf, Size:1020Kb
Load more
Recommended publications
-
Artificial Intelligence in Health Care: the Hope, the Hype, the Promise, the Peril
Artificial Intelligence in Health Care: The Hope, the Hype, the Promise, the Peril Michael Matheny, Sonoo Thadaney Israni, Mahnoor Ahmed, and Danielle Whicher, Editors WASHINGTON, DC NAM.EDU PREPUBLICATION COPY - Uncorrected Proofs NATIONAL ACADEMY OF MEDICINE • 500 Fifth Street, NW • WASHINGTON, DC 20001 NOTICE: This publication has undergone peer review according to procedures established by the National Academy of Medicine (NAM). Publication by the NAM worthy of public attention, but does not constitute endorsement of conclusions and recommendationssignifies that it is the by productthe NAM. of The a carefully views presented considered in processthis publication and is a contributionare those of individual contributors and do not represent formal consensus positions of the authors’ organizations; the NAM; or the National Academies of Sciences, Engineering, and Medicine. Library of Congress Cataloging-in-Publication Data to Come Copyright 2019 by the National Academy of Sciences. All rights reserved. Printed in the United States of America. Suggested citation: Matheny, M., S. Thadaney Israni, M. Ahmed, and D. Whicher, Editors. 2019. Artificial Intelligence in Health Care: The Hope, the Hype, the Promise, the Peril. NAM Special Publication. Washington, DC: National Academy of Medicine. PREPUBLICATION COPY - Uncorrected Proofs “Knowing is not enough; we must apply. Willing is not enough; we must do.” --GOETHE PREPUBLICATION COPY - Uncorrected Proofs ABOUT THE NATIONAL ACADEMY OF MEDICINE The National Academy of Medicine is one of three Academies constituting the Nation- al Academies of Sciences, Engineering, and Medicine (the National Academies). The Na- tional Academies provide independent, objective analysis and advice to the nation and conduct other activities to solve complex problems and inform public policy decisions. -
Artificial Intelligence
TechnoVision The Impact of AI 20 18 CONTENTS Foreword 3 Introduction 4 TechnoVision 2018 and Artificial Intelligence 5 Overview of TechnoVision 2018 14 Design for Digital 17 Invisible Infostructure 26 Applications Unleashed 33 Thriving on Data 40 Process on the Fly 47 You Experience 54 We Collaborate 61 Applying TechnoVision 68 Conclusion 77 2 TechnoVision 2018 The Impact of AI FOREWORD We introduce TechnoVision 2018, now in its eleventh year, with pride in reaching a second decade of providing a proven and relevant source of technology guidance and thought leadership to help enterprises navigate the compelling and yet complex opportunities for business. The longevity of this publication has been achieved through the insight of our colleagues, clients, and partners in applying TechnoVision as a technology guide and coupling that insight with the expert input of our authors and contributors who are highly engaged in monitoring and synthesizing the shift and emergence of technologies and the impact they have on enterprises globally. In this edition, we continue to build on the We believe that with TechnoVision 2018, you can framework that TechnoVision has established further crystalize your plans and bet on the right for several years with updates on last years’ technology disruptions, to continue to map and content, reflecting new insights, use cases, and traverse a successful digital journey. A journey technology solutions. which is not about a single destination, but rather a single mission to thrive in the digital epoch The featured main theme for TechnoVision 2018 through agile cycles of transformation delivering is AI, due to breakthroughs burgeoning the business outcomes. -
Congressional Record United States Th of America PROCEEDINGS and DEBATES of the 111 CONGRESS, FIRST SESSION
E PL UR UM IB N U U S Congressional Record United States th of America PROCEEDINGS AND DEBATES OF THE 111 CONGRESS, FIRST SESSION Vol. 155 WASHINGTON, WEDNESDAY, MARCH 4, 2009 No. 38 House of Representatives The House met at 10 a.m. and was May Your blessings be with those Ms. Potter has served in a number of called to order by the Speaker pro tem- suffering from the ravages of war and leadership roles at both the State and pore (Ms. JACKSON-LEE of Texas). our duty to them be ever on our minds. national levels of the American Legion f We are comforted by Your presence Auxiliary, and I would like to thank as we pray for a peaceful Nation. her personally for her ongoing service DESIGNATION OF THE SPEAKER In Your Name we pray, amen. to our Nation’s veterans. PRO TEMPORE f She is joined today by her husband, The SPEAKER pro tempore laid be- THE JOURNAL Toby, a retired Navy Seabee. fore the House the following commu- I ask my colleagues to join me in rec- nication from the Speaker: The SPEAKER pro tempore. The ognizing Ms. Potter for her service to WASHINGTON, DC, Chair has examined the Journal of the our country. March 4, 2009. last day’s proceedings and announces I hereby appoint the Honorable SHEILA to the House her approval thereof. f JACKSON-LEE to act as Speaker pro tempore Pursuant to clause 1, rule I, the Jour- ANNOUNCEMENT BY THE SPEAKER on this day. nal stands approved. NANCY PELOSI, PRO TEMPORE f Speaker of the House of Representatives. -
The Importance of Generation Order in Language Modeling
The Importance of Generation Order in Language Modeling Nicolas Ford∗ Daniel Duckworth Mohammad Norouzi George E. Dahl Google Brain fnicf,duckworthd,mnorouzi,[email protected] Abstract There has been interest in moving beyond the left-to-right generation order by developing alter- Neural language models are a critical compo- native multi-stage strategies such as syntax-aware nent of state-of-the-art systems for machine neural language models (Bowman et al., 2016) translation, summarization, audio transcrip- and latent variable models of text (Wood et al., tion, and other tasks. These language models are almost universally autoregressive in nature, 2011). Before embarking on a long-term research generating sentences one token at a time from program to find better generation strategies that left to right. This paper studies the influence of improve modern neural networks, one needs ev- token generation order on model quality via a idence that the generation strategy can make a novel two-pass language model that produces large difference. This paper presents one way of partially-filled sentence “templates” and then isolating the generation strategy from the general fills in missing tokens. We compare various neural network design problem. Our key techni- strategies for structuring these two passes and cal contribution involves developing a flexible and observe a surprisingly large variation in model quality. We find the most effective strategy tractable architecture that incorporates different generates function words in the first pass fol- generation orders, while enabling exact computa- lowed by content words in the second. We be- tion of the log-probabilities of a sentence. -
The Deep Learning Revolution and Its Implications for Computer Architecture and Chip Design
The Deep Learning Revolution and Its Implications for Computer Architecture and Chip Design Jeffrey Dean Google Research [email protected] Abstract The past decade has seen a remarkable series of advances in machine learning, and in particular deep learning approaches based on artificial neural networks, to improve our abilities to build more accurate systems across a broad range of areas, including computer vision, speech recognition, language translation, and natural language understanding tasks. This paper is a companion paper to a keynote talk at the 2020 International Solid-State Circuits Conference (ISSCC) discussing some of the advances in machine learning, and their implications on the kinds of computational devices we need to build, especially in the post-Moore’s Law-era. It also discusses some of the ways that machine learning may also be able to help with some aspects of the circuit design process. Finally, it provides a sketch of at least one interesting direction towards much larger-scale multi-task models that are sparsely activated and employ much more dynamic, example- and task-based routing than the machine learning models of today. Introduction The past decade has seen a remarkable series of advances in machine learning (ML), and in particular deep learning approaches based on artificial neural networks, to improve our abilities to build more accurate systems across a broad range of areas [LeCun et al. 2015]. Major areas of significant advances include computer vision [Krizhevsky et al. 2012, Szegedy et al. 2015, He et al. 2016, Real et al. 2017, Tan and Le 2019], speech recognition [Hinton et al. -
MS-042 John T. Reeder Photograph Collection
John T. Reeder Photograph Collection MS-042 Finding aid prepared by Elizabeth Russell, revised by Rachael Bussert. This finding aid was produced using the Archivists' Toolkit June 26, 2014 Describing Archives: A Content Standard Michigan Technological University Archives and Copper Country Historical Collections 1400 Townsend Drive Houghton, Michigan, 49931 906-487-2505 [email protected] John T. Reeder Photograph Collection MS-042 Table of Contents Summary Information ................................................................................................................................. 4 Biography.......................................................................................................................................................5 Collection Scope and Content Summary...................................................................................................... 5 Arrangement...................................................................................................................................................5 Administrative Information .........................................................................................................................6 Controlled Access Headings..........................................................................................................................6 Collection Inventory...................................................................................................................................... 8 Series I, Inventories and General Records..............................................................................................8 -
Spy Culture and the Making of the Modern Intelligence Agency: from Richard Hannay to James Bond to Drone Warfare By
Spy Culture and the Making of the Modern Intelligence Agency: From Richard Hannay to James Bond to Drone Warfare by Matthew A. Bellamy A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (English Language and Literature) in the University of Michigan 2018 Dissertation Committee: Associate Professor Susan Najita, Chair Professor Daniel Hack Professor Mika Lavaque-Manty Associate Professor Andrea Zemgulys Matthew A. Bellamy [email protected] ORCID iD: 0000-0001-6914-8116 © Matthew A. Bellamy 2018 DEDICATION This dissertation is dedicated to all my students, from those in Jacksonville, Florida to those in Port-au-Prince, Haiti and Ann Arbor, Michigan. It is also dedicated to the friends and mentors who have been with me over the seven years of my graduate career. Especially to Charity and Charisse. ii TABLE OF CONTENTS Dedication ii List of Figures v Abstract vi Chapter 1 Introduction: Espionage as the Loss of Agency 1 Methodology; or, Why Study Spy Fiction? 3 A Brief Overview of the Entwined Histories of Espionage as a Practice and Espionage as a Cultural Product 20 Chapter Outline: Chapters 2 and 3 31 Chapter Outline: Chapters 4, 5 and 6 40 Chapter 2 The Spy Agency as a Discursive Formation, Part 1: Conspiracy, Bureaucracy and the Espionage Mindset 52 The SPECTRE of the Many-Headed HYDRA: Conspiracy and the Public’s Experience of Spy Agencies 64 Writing in the Machine: Bureaucracy and Espionage 86 Chapter 3: The Spy Agency as a Discursive Formation, Part 2: Cruelty and Technophilia -
… … Mushi Production
1948 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 … Mushi Production (ancien) † / 1961 – 1973 Tezuka Productions / 1968 – Group TAC † / 1968 – 2010 Satelight / 1995 – GoHands / 2008 – 8-Bit / 2008 – Diomédéa / 2005 – Sunrise / 1971 – Deen / 1975 – Studio Kuma / 1977 – Studio Matrix / 2000 – Studio Dub / 1983 – Studio Takuranke / 1987 – Studio Gazelle / 1993 – Bones / 1998 – Kinema Citrus / 2008 – Lay-Duce / 2013 – Manglobe † / 2002 – 2015 Studio Bridge / 2007 – Bandai Namco Pictures / 2015 – Madhouse / 1972 – Triangle Staff † / 1987 – 2000 Studio Palm / 1999 – A.C.G.T. / 2000 – Nomad / 2003 – Studio Chizu / 2011 – MAPPA / 2011 – Studio Uni / 1972 – Tsuchida Pro † / 1976 – 1986 Studio Hibari / 1979 – Larx Entertainment / 2006 – Project No.9 / 2009 – Lerche / 2011 – Studio Fantasia / 1983 – 2016 Chaos Project / 1995 – Studio Comet / 1986 – Nakamura Production / 1974 – Shaft / 1975 – Studio Live / 1976 – Mushi Production (nouveau) / 1977 – A.P.P.P. / 1984 – Imagin / 1992 – Kyoto Animation / 1985 – Animation Do / 2000 – Ordet / 2007 – Mushi production 1948 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 … 1948 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 … Tatsunoko Production / 1962 – Ashi Production >> Production Reed / 1975 – Studio Plum / 1996/97 (?) – Actas / 1998 – I Move (アイムーヴ) / 2000 – Kaname Prod. -
In SLOW RECOVERY
THE PROFESSIONAL WINTER 2011 A Publication of the Associated Subcontractors of Massachusetts, Inc. Mass. Subcontractors Project of the Year – East Find BRIGHT SPOTS Genzyme in Project of the SLOW Year – West RECOVERY Springfield College Meet the New DCAM Commissioner Carole Cornelison THE PROFESSIONAL A Publication of the Associated Subcontractors of Massachusetts, Inc. cover story 16 Recovery is Slow, The Massachusetts economy and the construction industry appear to have entered But Some Sectors a recovery that’s been under way Are Booming Again for over a year. features 04 PRESIDENT’S VIEW 12 BIDDING Year in Review Understanding the Variables of Successful Bidding 06 BEACON HILL SPOTLIGHT Meet Carole Cornelison, 14 MARKETING AND PR DCAM Commissioner 3 Secrets Most Social Media Marketers Won’t Tell You 08 REO Court Invalidates Fall River 20 PROJECT OF THE YEAR – EAST Responsible Employer Genzyme Ordinance 22 PROJECT OF THE YEAR – WEST 10 INSURANCE Springfield College Purchasing Contractors’ Pollution Liability Coverage departments 24 PHOTO GALLERY Member Projects 2011 28 MEMBER NEWS The Professional Contractor 3 PRESIDENT’S VIEW BY DAVID G. CANNISTRARO Year in Review very year, this issue of The Professional Contractor fast-track rebuilding of Springfield College dorms devas- gives us an opportunity to look back at the year just tated by the June tornadoes. Both projects show project Epast, to comment on changes and showcase achieve- teamwork at its best. ments, and to look forward to the new year ahead. No recap of 2011 would be complete without men- This is now the fourth year this column has acknowl- tioning the one-year anniversary of the Massachusetts edged the down economy, and its major impact on Prompt Pay law, which took effect in November 2010. -
Evolutionary Design of Freecell Solvers Achiya Elyasaf, Ami Hauptman, and Moshe Sipper
IEEE TRANSACTIONS ON COMPUTATIONAL INTELLIGENCE AND AI IN GAMES 1 Evolutionary Design of FreeCell Solvers Achiya Elyasaf, Ami Hauptman, and Moshe Sipper Abstract—We evolve heuristics to guide staged deepening search for the hard game of FreeCell, obtaining top-notch solvers for this human-challenging puzzle. We first devise several novel heuristic measures using minimal domain knowledge and then use them as building blocks in two evolutionary setups involving a standard genetic algorithm and policy-based, genetic programming. Our evolved solvers outperform the best FreeCell solver to date by three distinct measures: 1) number of search nodes is reduced by over 78%; 2) time to solution is reduced by over 94%; and 3) average solution length is reduced by over 30%. Our top solver is the best published FreeCell player to date, solving 99.65% of the standard Microsoft 32K problem set. Moreover, it is able to convincingly beat high-ranking human players. Fig. 1. A FreeCell game configuration. Cascades: Bottom 8 piles. Foun- Index Terms—Evolutionary Algorithms, Genetic Algorithms, dations: 4 upper-right piles. Free cells: 4 upper-left cells. Note that cascades Genetic Programing, Heuristic, Hyper Heuristic, FreeCell are not arranged according to suits, but foundations are. Legal moves for the current configuration: 1) moving 7| from the leftmost cascade to either the pile fourth from the left (on top of the 8}), or to the pile third from the right I. INTRODUCTION (on top of the 8~); 2) moving the 6} from the right cascade to the left one (on top of the 7|); and 3) moving any single card on top of a cascade onto ISCRETE puzzles, also known as single-player games, the empty free cell. -
Solitaire: Man Versus Machine
Solitaire: Man Versus Machine Xiang Yan∗ Persi Diaconis∗ Paat Rusmevichientong† Benjamin Van Roy∗ ∗Stanford University {xyan,persi.diaconis,bvr}@stanford.edu †Cornell University [email protected] Abstract In this paper, we use the rollout method for policy improvement to an- alyze a version of Klondike solitaire. This version, sometimes called thoughtful solitaire, has all cards revealed to the player, but then follows the usual Klondike rules. A strategy that we establish, using iterated roll- outs, wins about twice as many games on average as an expert human player does. 1 Introduction Though proposed more than fifty years ago [1, 7], the effectiveness of the policy improve- ment algorithm remains a mystery. For discounted or average reward Markov decision problems with n states and two possible actions per state, the tightest known worst-case upper bound in terms of n on the number of iterations taken to find an optimal policy is O(2n/n) [9]. This is also the tightest known upper bound for deterministic Markov de- cision problems. It is surprising, however, that there are no known examples of Markov decision problems with two possible actions per state for which more than n + 2 iterations are required. A more intriguing fact is that even for problems with a large number of states – say, in the millions – an optimal policy is often delivered after only half a dozen or so iterations. In problems where n is enormous – say, a googol – this may appear to be a moot point because each iteration requires Ω(n) compute time. In particular, a policy is represented by a table with one action per state and each iteration improves the policy by updating each entry of this table. -
The Art of Insight in Science and Engineering
The Art of Insight in Science and Engineering 2014-09-02 10:51:35 UTC / rev 78ca0ee9dfae 2014-09-02 10:51:35 UTC / rev 78ca0ee9dfae The Art of Insight in Science and Engineering Mastering Complexity Sanjoy Mahajan The MIT Press Cambridge, Massachusetts London, England 2014-09-02 10:51:35 UTC / rev 78ca0ee9dfae © 2014 Sanjoy Mahajan The Art of Insight in Science and Engineering: Mastering Complexity by Sanjoy Mahajan (author) and MIT Press (publisher) is licensed under the Creative Commons At- tribution–Noncommercial–ShareAlike 4.0 International License. A copy of the license is available at creativecommons.org/licenses/by-nc-sa/4.0/ MIT Press books may be purchased at special quantity discounts for business or sales promotional use. For information, please email [email protected]. Typeset by the author in 10.5/13.3 Palatino and Computer Modern Sans using ConTEXt and LuaTEX. Library of Congress Cataloging-in-Publication Data Mahajan, Sanjoy, 1969- author. The art of insight in science and engineering : mastering complexity / Sanjoy Mahajan. pages cm Includes bibliographical references and index. ISBN 978-0-262-52654-8 (pbk. : alk. paper) 1. Statistical physics. 2. Estimation theory. 3. Hypothesis. 4. Problem solving. I. Title. QC174.85.E88M34 2014 501’.9-dc23 2014003652 Printed and bound in the United States of America 10 9 8 7 6 5 4 3 2 1 2014-09-02 10:51:35 UTC / rev 78ca0ee9dfae For my teachers, who showed me the way Peter Goldreich Carver Mead Sterl Phinney And for my students, one of whom said I used to be curious, naively curious.